Big Data and Analytics

Predictive AnalyticsBy Sarianne Gruber
Twitter: @subtleimpact

Michael Dulin, MD, PhD was a featured speaker at the Big Data Healthcare Analytics Forum, held in New York City in June. I had the pleasure to interview Dr. Dulin at the conference and learn about the Carolinas HealthCare System’s Dickson Advanced Analytics (DA2) group. They are leaders in Population Health boosting predictive analytics and making a huge impact on preventive care for patients and community health. Their state-of-the art Enterprise Data Warehouse integrates patient information from clinical, billing and claims data. This system brings together information from “more than 5 million patients into one repository that includes 1.5 petabytes of data across the system – equivalent to storing the entire print collection of the Library Congress”. Here are some of the first-of-its-kind programs that Dr. Dulin shared with me.

What you like to share about yourself and your current work the Carolinas HealthCare System?

I am the Chief Clinical Officer for Analytics and Outcomes Research at the Carolinas HealthCare System (CHS). It has been a real honor to be there over the past 17 years. And for the past 3 and a half years, I have been part of the analytics team, which we made major investments in building and developing an analytics and data competency within CHS. I think we really made some major strides as well. We are now seeing to some degree that we are seen as major leaders in the analytics field in healthcare domain. A lot of that is in our ability to bring data across multiple sources into one place to use. We brought clinical data, claims data, behavioral data, consumer data and all types of data together into one place. We started to build predictive models to better understand the risk of our patients. And to evaluate the outcomes of our interventions deployed across our systems as we look improve readmissions, quality of care delivery and rural population health. So it is a really exciting time for us. I think this going to be the future for healthcare, having this competency in data and analytics.

What is the magnitude of Carolinas Healthcare System organization?

We are one of the largest non-profit government healthcare systems in the country. We have right around 39 to 40 hospitals that are part of our system with over 600 care locations. We have somewhere in the range of 11 to 12 million patient encounters every year. And in terms of our transactions in our electronic medical record system, we have 20 million every day which is just a phenomenal amount of data that is being generated. When we started our process of building our enterprise data warehouse, we had somewhere in the range of 9 to 10 terabytes over the past year and half.  It has tripled to 30 terabytes of groomed data that sits in our enterprise data warehouse that we can use for analytics. We do take care of a lot of patients. We have a lot of encounters, and even small changes to improve efficiency or improve quality can impact 1000 of peopled lives. So that’s why sometimes people say, why are you as a primary care doctor in this data and analytics field? I really feel like what we are doing in this domain can impact more patients than I could ever impact as an individual physician seeing patients every day for the rest of my life. We really can impact lots of people’s lives.

Which clinical care areas at your hospitals are you focusing on with analytics?

I am particularly proud of, and I am primary care physician, so I am very excited about the work we are doing in the ambulatory domain. I am excited about the quality of care we provide for our patients with chronic conditions. For example, in diabetes, because of the entire package that we put together which includes data and analytics with the changes in the delivery system, we are now in the top decile in terms of our performance with patients with diabetes. And similarly with asthma, we have done a lot of work to improve care for patients and children with asthma.  With applications in analytics as well as shared decision making for interventions, we have been able to impact emergency department visits, hospitalizations, and really even improve quality of life for our patients with asthma. It has been really exciting for us in the ambulatory domain. In the acute care space, we also have done some really amazing work to try to impact readmissions and the length of stay for patients in the hospitals. We also try to reduce avoidable events like people having to go to ER department for something that is primary care treatable or preventable. So really just a lot of exciting work is going on.

One of the projects that we just completed was  forecasting our ER department volumes out 10 years  into the future. And for the first time, our system was really able to use data and understand our  entire population.  We looked over time to see what would happen as the population matured and try to understand what the population’s needs would be in terms of the emergency department to help make decisions about how we should design our healthcare system.

I understand you have taken very a creative approach to environmental health, air pollution control in helping your COPD and asthma patients. Can you tell us more about this initiative?

That has been a really fun project. I think we have a lot of opportunity to identify what can trigger asthma (in order) to change its course and keep our patients with asthma out of the hospital. One of the projects we did was to measure air pollution out in the community. We had our team members, our analysts and data scientists wear these monitors out in the community to get an idea about the pollution levels. We could help to match that (data) up and see where there were changes in the pollution levels that were triggering problems with our patients with asthma. Then we could identify those issues early on and warn our patients to maybe s to stay inside, stay out of certain areas so they won’t end up with problems with their asthma. That whole area of geospatial analysis for us I think is a really important competency. The ability to understand where patients live, what are the environmental factors are that can potentially influence their health outcomes, and then how do we improve the environment where they live so they can stay healthier, happier and out of our hospital system.

This is part I of our interview on the “impactfulness “of Big Data analytics.

About Michael Dulin, MD, PhD
Dr. Dulin helped to grow the Dickson Institute for Health Studies from a small research team into a centralized resource for data and analytics services for the system that employs over 100 employees. He was promoted into the position of Chief Clinical Officer for Analytics and Outcomes Research in 2013, where he oversees outcomes research for the system and co-leads the team. He received his undergraduate degree in electrical and biomedical engineering from the University of Texas. He earned his PhD from the University of Texas Health Science Center in Houston, and his medical degree from the University of Texas Medical School in Houston.

About Dickson Advanced Analytics (DA2) at Carolinas HealthCare System
The DA2 vision is to be an innovative leader in advanced analytics and business intelligence that predicts health needs, continually elevates patient outcomes, and drives drives transformative research and solutions to promote the health of our communities. Carolinas HealthCare System incorporates about 40 hospitals and over 600 care locations in North Carolina, South Carolina and Georgia. It is one of the largest HIT and EMR systems in the country. To learn more about  Dickson Advanced Analytics, visit the Carolinas HealthCare System website. Follow them on Twitter (@Carolinas).